Effects of Sediment Transport on Survival of Salmonid Embryos in a
Natural Stream: A Simulation Approach
Thomas E. Lisle and Jack Lewis
USDA Forest Service, Pacific Southwest Research Station, 1700 Bayview Drive, Arcata, CA 95521, USA
Lisle, T. E., and J. Lewis. 1992. Effects of sediment transport on survival of salmonid embryos in a natural stream:
a simulation approach. Can. J. Fish. Aquat. Sci. 49: 2337-2344.
A model is presented that simulates the effects of streamflow and sediment transport on survival of salmonid
embryos incubating in spawning gravels in a natural channel. Components of the model include a 6-yr streamflow
record, an empirical bed load-transport function, a relation between transport and infiltration of sandy bedload
into a gravel bed, effects of fine-sediment infiltration on gravel properties, and functions relating embryo survival
to gravel properties. High-flow events drive temporal variations in survival; cross-channel variations in bedload
transport cause spatial variations. Expected survival, as a result, varies widely from year to year and between
spawning runs in a single year. Alternative functions from previous research that relate survival to fine-sediment
concentration in spawning gravel and to intergravel rates of flow yield categorically different results. The relative
uncertainty of the components of this model indicates that the greatest research needs are to understand how
sediment transport affects the intergravel environment and how these changes affect embryo development and
survival.
Nous présentons un modèle qui simule les effets de l´écoulement d'un cours d'eau et du transport des sédiments
sur la survie des embryons de salmonidés en incubation clans des frayères de gravier dans un chenal naturel. Les
composantes du modèle incluent un relevé du débit du cours d'eau pendant 6 ans, une fonction empirique de
transport des sédiments de fond, une relation entre le transport et l´infiltration de la charge de fond sableuse dans
une couche de gravier, les effets de l´infiltration de sédiments fins sur les propriétés du gravier et des functions
mettant en relation la survie des embryons et les propriétés du gravier. Les épisodes de débit élevé entraînent
des variations temporelles de la survie; les variations D'travers le chenal du transport des sédiments de fond
causent des variations spatiales. Par conséquent, la survie prévue varie considérablement d'une année a l´autre
et d'une montaison de fraye à l'autre au cours de la même année. D'autres fonctions établies lors d´études
précédentes qui mettent en relation la survie et la concentration de sédiments fins clans le gravier de la frayère
ou l´écoulement entre les graviers donnent des résultats totalement différents. L´incertitude relative des composantes de ce modèle indique clue le besoin le plus pressant sur le plan de la recherche est la compréhension de
la manière dont le transport des sédiments influe sur l´environnement du gravier et des répercussions de ces
changements sur le développement et la survie des embryons.
Received January 2, 1992
Accepted May 21, 1992
(JB354)
urvival of salmonid embryos to emergence from the
streambed has been related to substrate and flow conditions in many experiments ( Phillips et al. 1975; Tappel
and Bjornn 1983; among others) and field studies (Koski 1966;
Tagart 1984; Scrivener and Brownlee 1989). The suitability of
incubation habitat ultimately depends on how much, what size,
and when sediment is transported. Almost regardless of the
original condition of gravel, the spawning female can alter the
grain size and porosity of gravel to ensure that the ova begin
with an adequate flow of oxygenated water (Burner 1951; Cordone and Kelley 1961; among others). For many regions and
species, however, incubation coincides with seasonal high
flows that carry sediment. Accumulated fine sediment in the
gravel can restrict intergravel flow and block emergence of fry.
Therefore, the key to embryo survival is not the condition of
spawning gravel before or immediately after spawning but during the several weeks or months of incubation. Moreover,
stream temperature can control the rate of development and
thereby determine the period of incubation (Heggberget 1988).
Processes that change gravel conditions vary widely in time and
space because they are driven by climatic events and modified
Can. J. Fish. Aquat. Sci., Vol. 49, 1992
Reçu le 2 janvier 1992
Accepté le 21 mai 1992
by the complexity of natural channels.
Are effects of sediment transport sufficiently intensive, pervasive, and frequent to significantly affect embryo survival?
To answer this question, we need to know how a hierarchy of
processes leading from (1) water discharge to (2) sediment
transport to (3) changes in gravel conditions to (4) physiologic
functions of embryos are linked. We also need to know how
the variability of processes affects survival of the population of
embryos incubating in a stream. This problem can be
approached effectively by modeling if relationships between
processes are quantifiable and data are available.
We present a model that links variations in flow and sediment
transport to the fraction of salmonids that survive in a natural
channel. We do not intend it to serve as a management tool,
but as a framework to build more accurate and comprehensive
models, reveal gaps in understanding, and explore the temporal
and spatial variability of sediment effects on embryo survival
in a natural stream. Furthermore, we do not presuppose that
quality of incubation habitat limits fish production. Rather, we
offer a mechanistic approach to deciphering the physical-biological linkage for one critical life stage of salmonids in streams.
2337
Study Site
We constructed the model with data primarily from one natural stream. Some site-specific relations are included pro tem
until more general relations can be found. We applied the model
to one reach where data are available; thus the model does not
evaluate embryo survival rates for the entire stream.
Jacoby Creek is a small gravel-bed stream that drains into
Humboldt Bay near Arcata, California. It supports populations
of steelhead trout (Oncorhynchus mykiss) and coho salmon
(O. kisutch) and has been the site of life-history studies of salmonids (Harper 1980), fluvial geomorphology (Lisle 1986),
and sediment transport and its effects on spawning gravels (Lisle
1989). Mean bankfull channel width at the primary study reach
is 17.2 m, median grain size of the bed surface is 22 nun, and
channel gradient is 0.0063. Drainage area is 36.3 km2, and
peak discharges can exceed 40 m3/s during the rainy season
from October through April. Clastic sediment yield (180 tons
(t)·km-2·yr-1), of which approximately 15% is bedload (Lehre
and Carver 1985), is moderate compared with other basins in
northwestern California (Janda and Nolan 1979). Large fluxes
of sediment during stormflow periods in Jacoby Creek can cause
extensive scour and fill and fine-sediment infiltration (Lisle
1989).
Model
The following provides a brief conceptual overview of our
model. We assume that adult fish are equally likely to enter
Jacoby Creek to spawn each day that discharge is between 0.85
and 4.8 m3/s during the period from January 1 to April 5. In
the process of burying the eggs, the females reduce the fraction
of fine sediment in the spawning gravel. The embryos then
incubate for 40 d, during which periodic high flows transport
fine sediment over the bed surface. Some sediment infiltrates
into the interstices of the gravel that overlies the embryos,
reducing intergravel flow velocities and plugging gravel interstices. These changes increase mortality of the embryos. Our
observations in Jacoby Creek suggest that redd topography is
planed off by flows large enough to transport significant quantities of sediment. We assume, therefore, that redd topography
does not affect infiltration of fine sediment.
Water Discharge
Water discharge drives changes in spawning gravels and is
the source of temporal variability of factors that affect embryo
survival. We used water stage, recorded continuously at a gauging station at the downstream end of the study reach, to
construct a record of discharge from 1979 to 1985. Discharge
during the rainy season ranged from 0.4 to 62 m3/s, and associated sediment transport ranged from a trace to hundreds of
tons per day. We used data for the period from January 1 to
May 15, which includes most of the rainy season. Calculations
were initiated after discharge first exceeded 0.85 m3/s each
winter, when it was assumed that adults entered the stream to
spawn.
Sediment Transport
Sand and granules (0.25 < D < 4 mm, where D is particle
diameter) are the predominant sizes that infiltrate spawning
gravels in Jacoby Creek (Lisle 1989). These particle sizes are
large enough to be transported in frequent contact with the bed
and small enough to enter interstices of spawning gravels.
2338
Water Discharge, m3/s
FIG. 1. Relation between transport rate of sand (0.25 < D < 2 mm)
and water discharge, Jacoby Creek.
Transport rates of this fine bedload were measured directly from
a bridge at the gauging station with a Helley-Smith bedload
sampler that had a mesh size of 0.20 mm. Fine-bedload transport rate increased steeply with discharge and had wide scatter
(Fig. 1) typical of bedload data from natural streams (Gomez
1991). As a point of reference, a transport rate of
0.023 kg·m-1.s-1 equates to 1 t·m-1.d-1.
We used a power function of transport rate versus water discharge to compute fine-bedload fluxes every hour during the
modeled period:
(1)
qb = (2.19 × 10-5)Q2.72
r2 = 0.71
where qb is mean bedload transport rate per unit channel width
(kilograms per metre per second) and Q is water discharge
(cubic metres per second). We computed mean unit flux of fine
bedload (kilograms per metre) by summing equation (1) for
each 40-d incubation period.
Infiltration of Fine Sediment
The volume of fine sediment that accumulates in spawning
gravels in Jacoby Creek is related to the flux of fine bedload
(Lisle 1989). Cans filled with gravel devoid of fine sediment
were buried in spawning areas to measure rates and grain sizes
of fine-sediment infiltration. After periods in which sediment
transport was measured continuously, we emptied the cans and
sieved their contents. For the model, infiltrated sediment was
considered to range between 0.12 and 4.0 mm in diameter,
which represents 93% of the material finer than 4 mm collected
in the cans. We considered this as the maximum size range that
we could use to determine infiltration rates from transport rates
of sediment between 0.25 and 4 mm.
Sediment infiltration rates were related to bedload transport
rates by a power function (Lisle 1989) modified to include a
wider range of grain size than used originally:
(2) I = 2.03(qB)T0.412
where I is mean fine-sediment infiltration (kilograms per square
metre of spawning gravel) and (qB)T is fine-bedload flux per
unit width (kilograms per metre). Data from two other streams
were also used to derive the original function (Lisle 1989). All
three streams have unimodal bed material and each shows no
systematic deviation from the common relationship between
Can. J. Fish. Aquat. Sci., Vol. 49, 1992
Relative fine-sediment infiltration
FIG. 2. Frequency distribution of relative fine-sediment infiltration
rates (fine-sediment content in each gravel can divided by the mean
of all cans for corresponding measurement periods) during stormflows, Jacoby Creek.
Particle Size, mm
FIG. 4. Cumulative size distributions by weight of fine-sediment infiltrate, bedload, spawning gravel whose interstices are completely filled
with fine sediment, and spawning gravel immediately after spawning,
Jacoby Creek.
recovered from five transects after six measured stormflow
periods. Values of local flux were normalized by dividing by
the mean value of flux for each measurement period, and the
probability density distribution of normalized fluxes was fit to
a gamma distribution (Fig. 3):
(3) ƒ [( q B )T ] = 5.351[( q B )T ]1.232 e −2.232( q B )T .
The discharge record and equations (1) through (3) could be
used to generate local sediment fluxes and infiltration rates for
each incubation period.
Changes in Spawning Gravel Conditions
Relative bedload flux
FIG. 3. Frequency and probability density distribution of relative bedload fluxes (local flux divided by mean flux), Jacoby Creek.
sediment infiltration and transport. Equation (2) was used to
compute fine-sediment infiltration for each incubation period.
Fine-sediment infiltration into the gravel cans varied widely
within each stormflow period (Fig. 2). To explicitly incorporate this variation in the model, we assumed that it was caused
solely by spatial variation in bedload transport. Bedload transport characteristically varies widely across channels (Carey
1985; Emmett, cited in Gomez 1991; Gomez et al. 1991).
Equation (2) indicates that infiltration rate for a given sediment
transport rate decreases as total sediment flux increases. The
physical explanation is that surficial interstices are free of fine
sediment during initial stages of infiltration and become plugged
as infiltration progresses, thereby inhibiting further infiltration.
A greater proportion of transported sediment can infiltrate areas
of the bed with low accumulated sediment fluxes more readily
than areas where high fluxes have previously plugged many
surficial interstices. This tends to dampen differences between
areas of the streambed in total sediment infiltrated.
We inverted equation (2) to compute local bedload fluxes
from values of infiltration measured in 153 cans that were
Can. J. Fish. Aquat. Sci., Vol. 49, 1992
We modeled effects of fine-sediment infiltration on gravel
properties that decrease embryo survival by two alternative
approaches: (1) an increase in the fraction of fine sediment in
the gravel, which affects both the rate of flow of oxygenated
water to the embryos and the ability of emerging fry to penetrate
the gravel interstices, and (2) a reduction in permeability of the
gravel, which decreases intergravel flow velocities.
In both approaches, we start with the average grain size distribution of Jacoby Creek spawning gravels, which contain a
small residual fraction of fine sediment, and recalculate gravel
properties at the end of each incubation period. We do not,
therefore, distinguish between adverse conditions that may prevail throughout most of an incubation period and those that may
accrue only near the end of the period. The particle-size distribution of gravel overlying freshly spawned eggs is the average distribution obtained from freeze-core samples at five
spawning areas (Lisle 1989) minus a proportion of the fraction
finer than 4 mm such that the fraction finer than 1 mm equals
10% (Fig. 4). Initial porosity of the gravel is 0.35. The thickness of the gravel layer is 0.2 m, which approximates the average burial depth of eggs of steelhead trout and coho salmon
(Shapalov and Taft 1954; Van den Berghe and Gross 1984).
The framework of the gravel bed after spawning is assumed
to remain intact as fine sediment fills the interstices. That is,
neither scour and fill nor expansion of the gravel framework
occurs. Particle-size distribution of fine sediment between 0.12
and 4 mm was obtained from the average distribution of material collected in the cans (Fig. 4). The disregarded fraction finer
2339
than 0.12 mm constituted only 7% of the material finer than
4 mm. Particle-size distribution of bedload material does not
equal that of infiltrated fines, partly because the distributions
are not truncated at the same minimum values. Also, fine bedload particles can more easily enter gravel interstices than coarse
particles and are thus disproportionately represented in sediment infiltrating the gravel (Einstein 1970; Carling 1984; Lisle
1989).
The introduction of fines changes the value of grain-size
parameters used in the model. In approach 1, bed-material fractions finer than 1 mm and finer than 8 mm were calculated for
increasing values of accumulation. These grain-size ranges correspond approximately to those used by Tappel and Bjornn
(1983) to relate survival to emergence of four species of
salmonids.
In approach 2, effective diameter, which is used in calculations of permeability, is reduced by fine-sediment infiltration.
Effective diameter is defined by
(4)
D e = [sum( p i / Ψi Di ]
−1
where pi is the volume fraction and Ψi is the sphericity of particles of size Di (Johnson 1980). Sphericity is given a value of
0.7.
Porosity is also reduced as interstices are filled with sediment. Given initial conditions, porosity (E) is calculated as a
function of fine-sediment infiltration (I, kilograms per square
metre):
(5)
E = E0—E0(1 - Ef)(I/Imax)
where E0 = 0.35 is initial porosity, Ef = 0.40 is porosity of
fine sediment, and Imax. = 0.042 kg/m2 is fine-sediment
accumulation to the point where interstices are completely
filled. Mass of sediment was converted to volume by dividing
by a particle density of 2.5 g/cm3 that was measured from
Jacoby Creek sediment. Using these values, equation (5) simplifies to
(6)
E = 0.35-5.0I.
E = 0.14 when interstices are completely filled. This value
corresponds closely to those for undisturbed gravel beds measured by Carling and Reader (1982).
Permeability (K) is calculated from the Karman-Cozeny
equation (Scheidegger 1960):
(7)
K = gf(E)De2/36 κν
where K is in centimetres per second, g is gravitational acceleration, f(E) = E3/(1 - E)2, and ν is kinematic viscosity. κ is
an empirically derived permeability constant and is given a
value of 6.4, which Johnson (1980) found suitable for spawning
gravel.
Apparent intergravel flow velocity (ua) is calculated from
Darcy's Law:
(8)
FIG. 5. Cumulative frequency distributions of particle sizes used in
our model and in experiments by Tappel and Bjomn (1983).
ua = K(dh/dl)
In the first approach, we use Tappel and Bjornn's (1983)
empirical relation between survival of steelhead embryos and the
percentages of spawning gravel finer than 0.85 mm and
finer than 9.5 mm:
(9A)
S = 94.7 - 0.1P9.5P0.85 + (P9.5)2
where S is expected percent survival to emergence and P9.5 and
P0.85 are the percent volume of sediment finer than 9.5 and
0.85 mm, respectively, in the overlying gravel. We substitute
P8 for P9.5 and P1 for P0.85. Although Tappell and Bjornn's
(1983) empirical relation fits specific laboratory conditions, we
justify its use in our model by comparing their gravel properties
with those used in our model. First, we both use a burial depth
of 0.2 m. Second, their particle-size distributions of gravelsand mixtures overlap our starting distribution and the distribution corresponding to maximum infiltration (Fig. 5).
Expected survival computed for a range of values of finesediment infiltration fits a linear relation well:
(9B)
S = 67.9 - 0.648I
r2 = 0.999
which, when combined with equation (2), yields
(9C)
S = 67.9 - 1.10(qB)T 0.412.
In our second approach, survival is a function of intergravel
flow velocity according to an empirical relation derived from
data for sockeye salmon (J. Pyper, cited in Cooper 1965):
r2 = 0.96.
(10A) S = 167.0 + 46.3log(ua)
The wall thickness and diameter of steelhead eggs are usually
smaller than those of sockeye salmon; thus, demands for oxygen concentration for steelhead should be greater (Wickett
1975; Beacham and Murray 1989). Consequently, equation
(10A) should overestimate survival of steelhead embryos.
Survival then approximates a linear function of infiltration
which is substituted for ua. using equations (4), (6), (7), and (8):
where ua is in centimetres per second. Head loss (dh/dl) over
riffle crests where fish commonly spawn is given a value of
0.005. Channel gradient of the study reach equals 0.0062. We
use values of ua to compute expected survival.
(10B) S=56.8 - 0.918I
Embryo Survival
For a given sediment flux, the survival function generated from
the Tappel and Bjornn (1983) relation (9C) yields higher rates
of survival than that generated from the intergravel-flow function (10C) (Fig. 6). This discrepancy is especially large for
We use the two approaches for evaluating changes in gravel
properties (described above) and corresponding relations to
embryo survival (described below) to explicitly evaluate uncertainty in relating survival to gravel properties.
2340
r2=0.999.
Then by using equation (2) to substitute (qB)T, for I:
(10C) S = 56.8 - 1.864(qB)T 0.412.
Can. J. Fish. Aquat. Sci., Vol. 49, 1992
FIG. 6. Functions of embryo survival versus bedload flux (equations
(9C) and (10C)) using the intergravel flow function and Tappel and
Bjornn's (1983) relation.
FIG. 7. Probability density distribution of embryo survival for all
40-d incubation periods over a 6-yr period using the intergravel-flow
function and Tappel and Bjornn's (1983) relation.
fluxes between 5000 and 20 000 kg/m, a range that was commonly
achieved during the 40-d periods.
Because each 40-d summation of mean sediment flux from
equation (1) is associated with a distribution of values of local
flux, expected survival must be integrated over equation (3)
using equations (9C) and (10C). This results in a distribution
of survival rates for each value of mean flux.
Results
Probability density distributions of expected embryo survival
were compiled for the entire period of record using the TappelBjornn and intergravel-flow functions (Fig. 7). Expected survival for each 40-d incubation period is the mean value resulting
from mean sediment flux and spatial variations in sediment flux
and infiltration. Variations in expected survival arise from temporal variation in sediment transport due to climatic events. The
salient feature of both distributions is the high variability of
survival rates.
The two survival functions yield categorically different sur- vival
rates. Results from the Tappel-Bjornn function indicate
Can. J. Fish. Aquat. Sci., Vol. 49, 1992
FIG. 8. Integrated relation computed from the intergravel-flow function between expected embryo survival and mean bedload flux with
± 1 SD in survival about the mean due to spatial variations in finesediment transport and infiltration.
that survival rates for each day eggs are deposited would most
often exceed 10%. In contrast, most of the distribution of survival rates computed from the intergravel-flow function falls
below 5%. Depending on critical embryo survival rates necessary to maintain fish populations, one could conclude that the
quality of incubation habitat does or does not limit populations,
depending on the choice of survival functions.
Spatial variations in sediment transport and infiltration create
high variations in expected survival, particularly for high values of sediment flux. The standard deviation of expected survival, expressed as either an absolute value or a ratio to mean
expected survival (coefficient of variation), increases with
increasing mean sediment flux computed from the intergravelflow function (Fig. 8). The coefficient of variation is greatest
for mean values of flux exceeding approximately 5000 kg/m
and corresponding to mean survival rates of less than 10%.
Given that such low survival rates are realized most of the time
according to this function (Fig. 7), high spatial variations typify expected survival in Jacoby Creek.
Temporal variations in expected survival are also high.
Embryo survival rates can be expected to vary widely from year
to year because of the episodic nature and strong influence of
stormflows that punctuate the spawning and incubation season
(Fig. 9). In some years, e.g. 1983, high-flow events are frequent, while in others, e.g. 1985, only one event may occur.
Even in a relatively dry year, e.g. 1980, survival may be low
if all eggs are laid over a short period and a single high-flow
event happens to occur during incubation.
Discussion
Evaluation of Uncertainty and Variability
One of the most useful immediate results of modeling a
system of natural processes is an evaluation of the relative
uncertainty and variability of individual processes. As more
knowledge of a process is gained and uncertainty diminishes,
our awareness of its variability often increases. Among the
hierarchy of processes between sediment transport and salmonid
embryo survival, variability is high and uncertainty relatively
low in the driving processes (streamflow and sediment
transport), and uncertainty is high in the response processes
(changes in gravel conditions and embryo mortality).
2341
Days Since January 1
FIG. 9. Hydrographs and expected survival of embryos for the period of January 1 to May 15 for the
years 1979-85. Plots of survival correspond to the day of emergence of alevins after a 40-d incubation
period following spawning which occurs within discharge limits of 0.85 and 4.8 m3/s.
Streamflow is highly variable but measured routinely with
little uncertainty. Bedload transport varies approximately with
streamflow, and varies considerably laterally and longitudinally
during a stormflow event (see Gomez 1991). Bedload transport
rates can be measured directly with careful attention to sampling
strategies and limitations of samplers or predicted with an
expected accuracy of within an order of magnitude (Gomez and
Church 1989). Considerable investment in measurement and
modeling is required to quantify spatial variation in bedload
transport. Compared with its effects on gravel properties and
embryo survival, however, variability in bedload transport
outweighs its uncertainty.
2342
Uncertainty in effects of sediment transport on spawning
gravel is high because of a lack of understanding of the variation
of fine-sediment infiltration and scour and fill with hydraulic
and sedimentological properties of streams. Scour and fill of
streambeds can alter streambed composition as much as
sediment infiltration (Lisle 1989) and can also result in fatal
entrainment of embryos (Gangmark and Broad 1956). Although
mean depths of scour and fill in gravel-bed channels have been
modeled (Ziemer et al. 1991), such models are crude and do
not adequately incorporate spatial variations. Relations between
sediment transport and fine-sediment infiltration are either
empirical, and thus applicable only to channels similar to those
Can. J. Fish. Aquat. Sci., Vol. 49, 1992
where
relations
were
developed,
or
fit
stringent
sedimentological criteria. Where sand is abundant, rates of
infiltration are influenced by plugging of gravel interstices near
the bed surface early in the infiltration process, which inhibits
further infiltration into deeper layers (Beschta and Jackson
1979; Carling 1984; Lisle 1989; Diplas and Parker 1992).
Alonso et al. (1988) developed a model of sediment infiltration
and its effects on gravel properties in a channel containing
abundant silt and clay, but little sand. Under these conditions,
silt and clay are able to penetrate all but the smallest interstices
and thus fill beds from the bottom up (Einstein 1968).
Uncertainty in the depth of fine-sediment infiltration
complicates modeling of changes in gravel conditions and their
effect on embryo survival. Empirical relations between survival
of eggs to emergence and the volume of fine sediment deposited
in redds vary considerably depending on grain size distributions
of fines and redd material, distribution of fines with depth in
the bed, and other conditions imposed in the laboratory or in
the field (Everest et al. 1987; Chapman 1988). As a result,
there are no clear criteria to choose between relations for a
model such as ours.
Chapman (1988), for example, objected to the application of
relations developed under laboratory conditions to field
situations on the grounds that gravel conditions in natural egg
pockets are poorly understood and not replicated by the uniform
distributions of fines with depth imposed in laboratory
experiments. The implication, however, that the greater
permeability that has been observed in an egg pocket (Platts
and Penton 1980) translates to a greater rate of intergravel flow
than indicated by the material surrounding the pocket may also
be erroneous. Subsurface streamlines do converge into a
permeable volume, but only when inflow and outflow conduits
exist. Thus, intergravel flow through a permeable egg pocket
may be governed more by conditions in the surrounding material
than those in the pocket itself. As Chapman (1988) admonished,
the true nature of bed material structure associated with redds
and the effects of sedimentation on embryo survival in natural
channels are not well understood and need to be improved
before accurate, predictive modeling can proceed. The
categorically different results from our model resulting from
two applicable survival functions highlight this need.
The issue is further complicated by multiple causes of
mortality, as well as adaptations of ova and alevins that increase
survival under low rates of intergravel flow and low levels of
dissolved oxygen (reviewed in Everest et al. 1987). Even if
intergravel flow is adequate for embryo development, sand that
plugs near-surface interstices can prevent alevins from emerging
from the gravel (Koski 1966; Phillips et al. 1975). Degradation
of incubation conditions can also led to smaller size of alevins
(Silver et al. 1963; Shumway et al. 1964; Bams 1969), which
may decrease chances for survival of fry. The immediate cause
of mortality from reduced intergravel flow is the reduction of
oxygen available to the embryos. Oxygen concentration of
intergravel water does not necessarily depend on intergravel
flow velocity, but they usually correlate (Chapman 1988). Fine
sediment in Jacoby Creek gravel contains abundant fine
particulate organic matter. Subsurface samples of gravel whose
interstices were filled with fine sand and silt smelled of decaying
organics, suggesting that low oxygen concentrations are
associated with high concentrations of fine sediment.
Influence of Diversity and Size of Spawning Runs
The strong influence of individual stormflow events suggests that
young-of-the-year fish populations would vary less widely
Can. J. Fish. Aquat. Sci., Vol. 49, 1992
if spawning were spread out in time over the rainy season. This
is most likely to result from a diversity of species, races, and
genotypes, as well as high escapement. Steelhead, for example,
spawn late in the season during the onset of drier weather in
many streams, and their embryos would have an advantage if
high flows became less frequent. Reductions in diversity of
genotypes that lead to concentrated spawning over a short period
would seem to threaten fish populations because even in relatively dry years the annual recruitment could be threatened by
a single stormflow. The number of embryos surviving each year
to provide annual fry recruitment is more likely to affect the
dynamics of fish populations than the composite survival rate
over a period of years.
Similarly, wide spatial variations in sediment effects suggest
that a diversity of redd sites can promote stability in fry recruitment. If a large number of redds are constructed, the few located
in zones of low sediment flux and infiltration may produce
enough fry to compensate for high mortality elsewhere. Steelhead do not appear to select low-risk spawning sites in Jacoby
Creek but instead choose loose gravel of the appropriate size
often in the thalweg. Our measurements have shown that sediment transport and scour and fill are high in such areas (Lisle 1989).
Fewer redds, resulting from reduced size and diversity of
spawning runs, would lead to increased variability in embryo
survival and a greater risk of dangerously low populations even
in the most favorable of seasons. As populations of emerging
fry become less, embryo survival rates become critical for fish
populations if the available rearing habitat for juveniles is not
adequately filled. At the same time, spawning success becomes
less predictable on the basis of the quality of incubation habitat
for embryos.
Conclusions
Despite a considerable degree of uncertainty and inherent
variability, the effects of flow and sediment transport on salmonid embryo survival can be modeled. Spatial and temporal
variability of water discharge, sediment transport, and finesediment infiltration can be explicitly incorporated. Relations
between driving and response variables can be quantified to
various degrees of generality, precision, and accuracy. The
resulting model is sufficiently comprehensive to explore variability, highlight research needs, and perhaps evaluate relative
effects of changes in sediment transport regime.
Regardless of the uncertainties in many of the processes we
modeled, it is clear that the fraction of embryos surviving to
emerge from spawning gravels each year can be expected to be
extremely variable in streams like Jacoby Creek, where fish
spawn during the high-flow season and the channel bed is at
least moderately mobile. In these cases, and especially where
adult escapement is low (Everest et al. 1987), the nature of the
linkage between the sediment transport regime and embryo
mortality may be to limit fish production in some years but not
others.
Some research needs are clear. The further we move down
the hierarchy of processes from water discharge to embryo mortality the more we find is unknown. Relations between sediment transport and gravel conditions need to become more
general, include scour and fill, and adequately treat variations
in depth and grain size of fine-sediment infiltration. More crucially, the subsurface structure of natural redds and the response
of embryos to conditions in spawning gravels need to be quantified better. Progress in the problems outlined above will move
2343
us closer to predicting embryo survival under changing watershed
conditions and improve our understanding of factors
influencing population dynamics of stream fishes.
Acknowledgments
We thank Don Chapman, Ike Schlosser, and Charles Scrivener for
helpful reviews of our paper.
References
ALONSO, C. V., F. D. THEURER
AND D. W. ZACHMANN. 1988. Tucannon River
offsite study: sediment intrusion and dissolved-oxygen transport model.
USDA, Agricul. Res. Serv., Hydro-Ecosystems Res. Group, Fort Collins,
CO.
BAMS, R. A. 1969. Adaptions of sockeye salmon associated with incubation
in stream gravels, p. 71-87. In T. G. Northcote [ed.] Symposium on
salmon and trout in streams, H. R. MacMillan Lectures in Fisheries, University of British Columbia, Vancouver, B.C.
BEACHAM, T. D., AND C. B. MURRAY. 1989. Variation in developmental biology of sockeye salmon (Oncorhynchus nerka) and chinook salmon
(O. tshawytscha) in British Columbia. Can. J. Zool. 67: 2081-2089.
BESCHTA, R. L., AND W. L. JACKSON. 1979. The intrusion of fine sediments
into a stable gravel bed. J. Fish. Res. Board Can. 36: 204-210.
BURNER, C. J. 1951. Characteristics of spawning nests of Columbia River
salmon. U.S. Fish Wildl. Serv. Fish. Bull. 52(61): 97-110.
CAREY, W. P. 1985. Variability in measured bedload-transport rates. Water
Resour. Bull. 21(1 ): 39-48.
CARLING, P. A. 1984. Deposition of fine and coarse sand in an open-work
gravel bed. Can. J. Fish. Aquat. Sci. 41: 263-270.
CARLING, P. A., AND N. A. READER. 1982. Structure, composition and bulk
properties of upland stream gravels. Earth Surf. Proc. Landforms 7: 349365.
CHAPMAN, D. W. 1988. Critical review of variables used to define effects of
fines in redds of large salmonids. Trans. Am. Fish. Soc. 117(1): 1-21.
COOPER, A. C. 1965. The effect of transported stream sediments on the survival
of sockeye and pink salmon eggs and alevin. Int. Pac. Salmon Fish.
Comm. Bull. 17: 71 p.
CORDONE, A. J., AND D. W. KELLEY. 1961. The influences of inorganic sediment on the aquatic life of streams. Calif. Fish Game 47(2): 189-228.
DIPLAS P., AND G. PARKER. 1992. Deposition and removal of fines in gravelbed streams. In P. Billi, R. D. Hey, C. R. Thome, and P. Tacconi [ed.]
Dynamics of gravel-bed rivers. J. Wiley & Sons, Ltd., Chichester, U.K.
(In press)
EINSTEIN, H. A. 1968. Deposition of suspended particles in a gravel bed.
J. Hydraul. Div. ASCE 94(HY5): 1197-1205.
1970. Closure, Deposition of suspended particles in a gravel bed.
J. Hydraul. Div. ASCE 96(HY2): 581-582.
EVEREST, F. H., R. L. BESCHTA, J. D. SCRIVENER, K. V. KOSKI, J. R. SEDELL,
AND C. J. CEDERHOLM. 1987. Fine sediment and salmonid production—
a paradox, p. 98-142. In E. Salo and T. Cundy fed.] Streamside management and forestry and fishery interactions. University of Washington,
College of Forestry Resources, Cont. 57, Seattle, WA.
GANGMARK, H. A., AND R. D. BROAD. 1956. Further observations on stream
survival of king salmon spawn. Calif. Fish Game 42: 37-49.
GOMEZ, B. 1991. Bedload transport. Earth-Sci. Rev. 31: 89-132.
GOMEZ, B., AND M. CHURCH. 1989. An assessment of bed load sediment transport formulae for gravel bed rivers. Water Resour. Res. 25(6): 11611186.
GOMEZ, B., W. W. EMMETT, AND D. W. HUBBELL. 1991. Comments on sampling bedload in small rivers. Proc. 5th Fed. Inter-Agency Sediment. Conf.
1: 2.65-2.72.
2344
HARPER, W. G. 1980. Age, growth, and migration of coho salmon and steelhead trout in Jacoby Creek, California. M.S. thesis, Humboldt State University, Arcata, CA. 103 p.
HEGGBERGET, T. G. 1988. Timing of spawning in Norwegian Atlantic Salmon
(Salmo salar). Can. J. Fish. Aquat. Sci. 45: 845-849.
JANDA, R. A., AND D K. M. NOLAN. 1979. Stream sediment discharge in northwestern California, p. IV-1 to IV-27. In Natural and management-related
erosion in Franciscan Terrane of northern California. Geol. Soc. Am.
Cordilleran Sect. Guidebook, Field Trip 12. San Jose State University,
San Jose, CA.
JOHNSON, R. A. 1980. Oxygen transport in salmon spawning gravels. Can. J.
Fish. Aquat. Sci. 37: 155-162.
KOSKI, K. V. 1966. The survival of coho salmon (Oncorhynchus kisutch) from
egg deposition to emergence in three Oregon coastal streams. M.S. thesis,
Oregon State University, Corvallis, OR. 84 p.
LEHRE, A. K., AND G. CARVER. 1985. Thrust faulting and earthflows: speculations on the sediment budget of a tectonically active drainage basin,
p. 169-182. In M. E. Savina [ed.] Redwood Country, American Geomorphological Field Group 1985 Field Trip Guidebook, Humboldt State
University, Arcata, CA.
LISLE, T. E. 1986. Stabilization of a gravel channel by larve strearnside obstructions and bedrock bends, Jacoby Creek, northwestern California. Geol.
Soc. Am. Bull. 97: 999-1011.
1989. Sediment transport and resulting deposition in spawning gravels, north coastal California. Water Resour. Res. 25(6): 1303-1319.
PHILLIPS, R. W., R. L. LANTZ, E. W. CLAIRE, AND J. R. MORING. 1975. Some
effects of gravel mixtures on emergence of coho salmon and steelhead
trout fry. Trans. Am. Fish. Soc. 104(3): 461-466.
PLATTS, W. S., AND V. E. PENTON. 1980. A new freezing technique for sampling salmonid redds. USDA For. Serv. Res. Pap. INT-248: 22 p.
SCHEIDEGGER, A. E. 1960. The physics of flow through porous media. Macmillan Co., New York, NY. 313 p.
SCRIVENER, J. C., AND M. J. BROWNLEE. 1989. Effects of forest harvesting on
spawning gravel and incubation survival of chum (Oncorhynchus keta)
and coho salmon (O. kisutch) in Carnation Creek, British Columbia. Can.
J. Fish. Aquat. Sci. 46: 681-696.
SHAPALOV, L., AND A. C. TAFT. 1954. The life histories of the steelhead rainbow trout (Salmo gairdneri gairdneri) and silver salmon (Oncorhynchus
kisutch) with special reference to Waddell Creek, California, and recommendations regarding their management. Calif. Dep: Fish Game Bull.
98: 375 p.
SHUMWAY, D. L., D. E. WARREN, AND P. DOUDOROFF. 1964. Influence of
oxygen concentration and water movement on the growth of steelhead
trout and coho salmon embryos. Trans. Am. Fish. Soc. 93(4): 342-456.
SILVER, S. T., C. E. WARREN, AND P. DOUDOROFF. 1963. Dissolved oxygen
requirements of developing steelhead trout and chinook salmon embryos
at different water velocities. Trans. Am. Fish. Soc. 92(4): 327-343.
TAGART, J. V. 1984. Coho salmon survival from egg deposition to fry emergence, p. 173-181. In J. M. Walton and D. B. Houston [ed.] Proc. Olympic Wild Fish Conference. Peninsula College and Olympic National Park,
Port Angeles, WA.
TAPPEL, P. D., AND T. C. BJORNN. 1983. A new method of relating size of
spawning gravel to salmonid embryo survival. N. Am. J. Fish. Manage.
3: 123-135.
VAN DEN BERGHE, E. P., AND M. R. GROSS. 1984. Female size and nest depth
in coho salmon (Oncorhynchus kisutch). Can. J. Fish. Aquat. Sci. 41:
204-206.
WICKETT, W. P. 1975. Mass transfer theory and the culture of fish eggs. In
W. Adams [ed.] Trans. Electrochem. Soc. Princeton, NJ 419-433.
ZIEMER, R. R., J. LEWIS, R. M. RICE, AND T. E. LISLE. 1991. Modeling the
cumulative watershed effects of forest management strategies. J. Environ.
Qual. 20(1): 36-42.
Can. J. Fish. Aquat. Sci., Vol. 49, 1992